Exercise efficiency at low power outputs, energetically comparable to daily living activities, can be influenced by homeostatic perturbations (e.g., weight gain/loss). However, an appropriate efficiency calculation for low power outputs used in these studies has not been determined. Fifteen active subjects (seven females, eight males) performed 14, 5-min cycling trials: two types of seated rest (cranks vertical and horizontal), passive (motor-driven) cycling, no-chain cycling, no-load cycling, cycling at low (10, 20, 30, 40 W), and moderate (50, 60, 80, 100, 120 W) power outputs. Mean delta efficiency was 57% for low power outputs compared to 41.3% for moderate power outputs. Means for gross (3.6%) and net (5.7%) efficiencies were low at the lowest power output. At low power outputs, delta and work efficiency values exceeded theoretical values. In conclusion, at low power outputs, none of the common exercise efficiency calculations gave values comparable to theoretical muscle efficiency. However, gross efficiency and the slope and intercept of the metabolic power vs mechanical power output regression provide insights that are still valuable when studying homeostatic perturbations.